Numerical Weather Prediction • Reanalysis • Machine Learning

Atmospheric & AI Scientist for Earth System Data

I develop methods to extract robust, actionable information from weather and climate data – combining numerical models, reanalyses such as ERA5, and machine learning. My work spans from energy-relevant weather regimes and snowpack to quality assessment of global Earth observation products.

Current projects Scientific profile

About

I am a Senior Research Scholar in the Department of Geophysics at Boise State University, with a background in Earth system modeling, chemistry–climate interactions, and climate services for energy and risk. My current work focuses on combining numerical models, reanalyses and AI methods to better understand weather and climate variability.

Profile

I hold a Ph.D. in Physics and have 15+ years of experience in atmospheric research – from stratospheric ozone and chemistry–climate modeling to the evaluation of CMIP5/CMIP6 Earth System Models and reanalyses such as ERA5. I have contributed to community tools for model evaluation (ESMValTool, CCMValTool) and to several international initiatives.

My expertise sits at the interface between physical understanding and data science: I enjoy building diagnostics, workflows and machine learning approaches that turn large, heterogeneous datasets into reliable, useful information for decision-making.

Current position

  • Senior Research Scholar, Department of Geophysics, Boise State University.
  • Machine learning analysis of Perturbed Parameter Ensembles to evaluate soil and root parameters in land-surface and hydrological models.
  • Integration of land–atmosphere processes, renewable-energy relevant variables, and high-resolution modeling to support energy and water applications in complex terrain.

Research Themes

My work combines numerical weather and climate models, reanalysis datasets such as ERA5, and machine learning to better understand variability, extremes and their implications for energy and hydrology.

Energy-relevant weather regimes

Characterising multi-day low-wind and low-solar “energy drought” events (Dunkelflaute) using CESM2 ensembles, reanalyses and clustering / ML methods, with applications to grid reliability and renewable integration.

High-resolution snowpack & water

Downscaling Snow Water Equivalent (SWE) from Earth System Models using deep learning, calibrated against NLDAS-3, to deliver actionable information for mountain hydrology and western U.S. water resources.

Model evaluation & diagnostics

Long-standing experience in evaluating CMIP5/CMIP6 and chemistry–climate models, contributing to community tools (ESMValTool, CCMValTool) and to the Copernicus Climate Data Store independent quality assessment.

Selected Projects

A selection of ongoing and recent work at the intersection of atmospheric science, numerical weather and climate models, and machine learning – with a focus on energy, water, and Earth observation data.

Compound Energy Drought in Idaho & the Pacific Northwest

NSF I-CREWS Seed – Convergence Research

Analysis of multi-day low-wind and low-solar events (Dunkelflaute) using CESM2-LE and reanalysis data to quantify risks for wind and solar production. The project combines weather regime analysis, clustering and ML to understand drivers and recurrence patterns, with a proof-of-concept application to the planned Lava Ridge wind project.

Deep Learning for High-Resolution SWE Projections

NASA EPSCoR Concept

Development of deep learning models to downscale Snow Water Equivalent from coarse-resolution Earth System Models to hydrologically relevant scales, calibrated and evaluated using NLDAS-3 and detailed topographic information across mountain regions in the western U.S.

Copernicus Climate Data Store – ECV Quality Assessment

Copernicus C3S

Principal Investigator for the independent quality assessment of Essential Climate Variables (satellite, in situ and reanalysis), assessing reliability and fitness-for-purpose for a wide range of climate applications, including energy and water management.

Climate Services for the Energy Sector

Hybrid Statistical Modeling

Development of hybrid methods combining seasonal forecasts and observations to improve predictions of energy-relevant climate variables over Europe, integrating statistical modeling, teleconnection indices and model output to support decision-making in the power sector.

Publications

Author and co-author of 40+ peer-reviewed articles in atmospheric science, climate and Earth system modeling. A full, up-to-date list is available via ORCID.

Highlighted topics

  • Stratospheric ozone response to geoengineering and volcanic forcing.
  • Evaluation of CMIP5/CMIP6 Earth system models and chemistry–climate models.
  • Seasonal predictions of energy-relevant variables using hybrid statistical approaches.
  • Diagnostics and metrics for large multi-model ensembles.

Full publication list: ORCID 0000-0002-0591-9193

Contact

For collaborations, speaking engagements, or questions about my work, feel free to get in touch.